The randomness of simulation: a test - General Discussion - FM24/25 Update

Fantastic

14 years ago

9 months ago

2,496

Banned

02 May 2014 03:18:06

I'll start this discussion by offering an amnesty: we have all saved and reloaded. Cheating in Football Manager, specifically those of us who are members of the online community, is a massive faux pas, but any person who wants to sit there and claim they have never saved and reloaded in order to receive a favourable result is a liar. Plain and simple.

That rant is relevant for the topic I want to discuss today (and test): how far reaching is the randomness of such a complex simulation?

To make it easier to understand I'll analogise it: how many times in our careers have we been unable to pinpoint why we don't win, why we're not getting results against supposedly inferior opposition? Far too often to test our sanity. The genetic makeup of Football Manger (and football itself) is that on any given day any team can beat any other team. It's why we have examples of ridiculously underpowered teams toppling top flight opposition. It's part and parcel of what makes this sport exciting for people.

This is inspired by me losing to a team far worse than I am and not understanding why. I lost, yet the week before I beat a team far better than I am. I'm confident the players I have are good enough, the tactics (as basic as they are) are good, the players I'm using are motivated... so why did I lose? Is it random, is it design, is it logic?

So what I'm proposing is embracing something that people have been doing for eons: a test that takes into account saving and reloading.I want to take over as manager of a team in the Premier League and just use one season as the empirical evidence. But what I will do is take a decent tactic, nominated by the members here, and play every game during the season ten times each. So at the end I will have played 380 matches. Then using the magic of maths I can work out the absolute worst season-length points tally I could have, the best, and the average. Just to give a little insight into this game just being a series of interspersing moments can affect the overall outcome.

So what I need from you guys is a team you think it would be worthwhile to test this out with (hint: not one that is going to win 75% of games) and a tactic that people consider to be solid. Of course any other suggestions/comments are greatly appreciated.

.verse

13 years ago

1 year ago

871

02 May 2014 08:42:00

I think a side like Newcastle, Aston Villa, Everton, W.B.A. or Southampton would be the most suitable EPL sides.

Tactically, you'll want to go with a rather common place tactic that is widely used, balanced and proven: 4-3-2-1 (DM-CMx2, AML/R), the rather popular 4-2-3-1 (CMx2, AML/C/R), or a simple 4-4-2.

I did a bit of a test (FMC mode) with this during a LLM game I felt I was being screwed out of. I replayed the game a handful of times combating the exact problem areas required, the games played largely different from the initial one, but the result's still weren't reflective of this. There was nothing I could seemingly do though that would provide a win, with the result either being a draw or the original result of a loss. Keep in mind at the time my side was pushing for a promotion place and the opposition were dead last. In the end I tried using the instant result feature without any luck either initially, until finally getting a win after about the 4th or 5th attempt.

I've always felt that there are some games that have results somewhat predetermined, no matter how well you counter the opposition. They'll always score that jammy free kick or corner.

KEZ_7

18 years ago

2 weeks ago

1,884

02 May 2014 09:04:46

Like the idea of this. I'd take Newcastle, not for bias but for the fact that I've seen them finish anywhere from 3rd to 19th under the AI so the potential is there for it to go either way. Try a formation with a diamond midfield, seems to be the favourite of the scene at the moment

Sheriff Skacel

10 years ago

9 months ago

940

Premium

02 May 2014 10:25:46

I ran a test with tactics earlier this year, I took over as manager at a host of clubs, gave them all the same tactics and holidayed a season. I got sacked from the likes of Arsenal and Liverpool, nearly getting Arsenal relegated, I won the EPL with Newcastle, I got sacked from Barcelona but won La Liga with Real Madrid, can't remember what happened in Serie A.

I felt like sharing this.

Poe

18 years ago

1 week ago

3,675

02 May 2014 11:09:48

I feel the match engine is increasingly random.

I really don't understand the FM14 match engine.

BOT Brad

11 years ago

5 years ago

166

02 May 2014 12:21:14

I think a team like Southampton or Newcastle would be a good fit for this kind of test. I have seen both Southampton and Newcastle relegated in various saves in the 2013-14 season, but also have seen Southampton win the league and Newcastle finished 2nd or 3rd a few times too.